¿du
A novel mechanism of tumorigenesis involving pH-dependent destabilization of a mutant p53 tetramer
Enrico L. DiGiammarino1, Amanda S. Lee1, Craig Cadwell2, Weixing Zhang1, Brian Bothner1,3, Raul C. Ribeiro4,5, Gerard Zambetti2,3 and Richard W. Kriwacki1,3
Departments of 1Structural Biology and 2Biochemistry, St. Jude Children’s Research Hospital, 332 N. Lauderdale St., Memphis, Tennessee 38105, USA. 3Department of Molecular Sciences, University of Tennessee Health Sciences Center, Memphis, Tennessee 38163, USA. 4Department of Hematology/ Oncology and 5 International Outreach Program, St. Jude Children’s Research Hospital, 332 N. Lauderdale St., Memphis, Tennessee 38105, USA.
Published online: 26 November 2001, DOI: 10.1038/nsb730
The p53 tumor suppressor requires tetramerization to func- tion as an initiator of cell cycle arrest and/or apoptosis. Children in southern Brazil that exhibit an elevated incidence of adrenocortical carcinoma (ACC) harbor an Arg 337 to His mutation within the tetramerization domain of p53 (p53-R337H; 35 of 36 patients). The mutant tetramerization domain (p53tet-R337H) adopts a native-like fold but is less sta- ble than the wild type domain (p53tet-wt). Furthermore, the stability of p53tet-R337H is highly sensitive to pH in the physi- ological range; this sensitivity correlates with the protonation state of the mutated His 337. These results demonstrate a pH- sensitive molecular defect of p53 (R337H), suggesting that pH- dependent p53 dysfunction is the molecular basis for these cases of ACC in Brazilian children.
DNA damage and hyperproliferative signals activate the tumor suppressor p53, leading to either cell cycle arrest or apop- tosis1,2. In this manner, p53 monitors genome integrity and, when necessary, activates mechanisms that allow DNA damage to be repaired or damaged cells to be eliminated via apoptosis. Through its essential role in genome surveillance, inactivation of p53 commonly results in the formation of tumors. For example, >50% of all human cancers have mutations in p53 (ref. 3). Human p53 is comprised of four domains: an N-terminal trans- activation domain, a central DNA binding domain, a tetramer- ization domain and a C-terminal basic domain. The DNA binding domain and the tetramerization domain are discretely folded, independent domains4; more important, tetramerization is required for tumor suppression activity5. The majority of can- cer-associated p53 mutations cluster within the DNA binding domain and disrupt DNA binding activity3. However, recent studies have identified mutations within the tetramerization domain that are associated with cancer6-9.
The structure of the p53 tetramerization domain, contained within residues 310-360, has been determined by both NMR spectroscopy10,11 and X-ray crystallography12. This domain forms a compact structure that is a dimer of dimers (Fig. 1b,c), with a four-helix bundle flanked by antiparallel ß-strands. The dimer-dimer interface is stabilized by hydrophobic interactions, whereas the individual dimers are stabilized by hydrophobic and
a
Position 337
H
310
1
360
GSHM NNTSSSPQPKKKPLDGEYFTLQIRGRERFEMFRELNEALELKDAQAGKEPG
ß-strand
a-helix
b
c
Arg 337
Asp 352
ionic interactions between the helices and hydrogen bonds between the ß-strands. The wild type domain (p53tet-wt) forms stable tetramers under physiological conditions and can be reversibly unfolded13.
Because of the inherent structural symmetry of this domain, point mutations can significantly reduce its stability. For exam- ple, mutation of Leu 344 to Pro (p53-L344P) is associated with Li-Fraumeni syndrome (LFS) in which a broad spectrum of tumors occurs, including brain tumors, sarcomas, breast tumors, leukemias and adrenal cortical tumors6. This mutation disrupts tetramerization and sequence-specific DNA binding8. Another inherited mutation within the tetramerization domain, Arg 337 to Cys, is associated with Li-Fraumeni-like syndrome (LFLS)7. Arg 337 forms a salt bridge with Asp 352 across the helix-helix interface within dimer subunits12 (Fig. 1b,c); four such interactions are observed in the tetramer. The R337C mutation decreases the midpoint of the thermal unfolding curve for the tetramerization domain by 52 ℃, from 85 ℃ to 33 ℃, causing the domain to be mostly unfolded under physiological conditions. This extent of structural destabilization of the tetramerization domain is associated with a predisposition to cancer8.
Recently, mutation of Arg 337 to His (p53-R337H) was linked to pediatric adrenal cortical carcinoma (ACC)9. Ribeiro and coworkers9 report that in a localized population in southern Brazil, 97% of children (n = 36) that developed tumors in the adrenal gland harbored the R337H mutation of p53. Although a broad spectrum of cancer types, including ACC, is often associ- ated with inheritance of mutated p53, only pediatric ACC is observed in patients with the p53-R337H genotype. In contrast to the disruption of function caused by other cancer-linked p53 mutations, two groups have reported that the function of
npg
a
b
[O]MR X 10-3 (deg x cm2 x dmol-1)
4
2
p53tet-wt
e
280 nm
1
MW standards
0
p53tet-R337H
c
d
p53tet-wt
-2
Absorbance
b
p53tet-R337H
-4
-6
a
-8
-10
-12
0
-14
0
Retention Time (min)
20
40
60
190
200
Wavelength (nm)
210
220
230
240
250
260
c
d
110.0
110.0
115.0
15N (ppm)
115.0
15N (ppm)
120.0
120.0
125.0
125.0
9.5
9.0
8.5
8.0
7.5
7.0
6.5
9.5
9.0
8.5
8.0
7.5
7.0
6.5
1H (ppm)
1H (ppm)
p53-R337H is indistinguishable from that of p53-wt9,14. These results seem to be in conflict with the clear evidence linking this mutation with pediatric ACC. To resolve this discrepancy, we investigated the effects of this mutation on the structure and sta- bility of the p53 tetramerization domain.
p53tet-R337H and p53tet-wt are structurally similar We compared the structures of p53tet-wt and p53tet-R337H using several methods. First, gel filtration chromatography and glutaraldehyde crosslinking were used to monitor the oligomeric state. p53tet-R337H and p53tet-wt elute as tetramers from a gel filtration column (Fig. 2a) and have simi- lar crosslinking profiles when treated with glutaraldehyde (data not shown). These data establish that both p53tet-R337H and p53tet-wt are tetramers. Second, circular dichroism (CD) showed that the two domains have similar secondary structure (Fig. 2b). Finally, 2D 1H-15N TROSY spectra15 were used to qualitatively compare the structures of the two tetramerization domains (Fig. 2c,d). Although small differences in peak posi- tions are observed between the two spectra, the overall patterns of crosspeaks are very similar. Collectively, these data indicate that, although not identical, p53tet-R337H and p53tet-wt are structurally very similar.
Thermal stability of p53tet-R337H
CD was used to determine the thermal stability of p53tet-
R337H versus p53tet-wt over a range of pH values (Fig. 3).
p53tet-R337H unfolds at significantly lower temperatures than
p53tet-wt at all pH values, and its stability is very pH dependent
(Fig. 3a). In contrast, p53tet-wt is highly stable at all pH values
tested (Fig. 3b). At pH 5 and 6, p53tet-R337H denatures at
52 ℃ (Fig. 3a, red and yellow traces, respectively). Although
these Tm values are20 ℃ lower than those for p53tet-wt deter-
mined here (Fig. 3b, red and yellow traces, respectively), <10%
of p53tet-R337H molecules are unfolded at 37 ℃. However,
p53tet-R337H becomes significantly less stable as pH is
increased toward the physiological range between 7 and 8
(Fig. 3a, green and cyan traces, respectively). For p53tet-R337H
at pH 8 and 37 ℃, the fraction of unfolded tetramerization
domain molecules is substantial (~70% unfolded). In contrast,
p53tet-wt is fully folded under the same conditions. At pH 7
and 37 ℃, only ~20% of p53tet-R337H molecules are unfolded,
highlighting the steep pH dependence of stability. Therefore,
the R337H mutation found in patients with ACC greatly desta-
bilizes the p53 tetramerization domain at the high end of the
physiological pH range (generally considered to be between
pH 6.5 and 8.0) 16.
npg
Fig. 3 p53tet-R337H is destabilized in a pH-dependent manner. a, The thermal stability of p53tet-R337H, associated with ACC in Brazilian chil- dren, is reduced with respect to that of b, p53tet-wt. CD ellipticity at 222 nm was measured, and these values were used to calculate the mole fraction of denatured molecules (fu) at each temperature (15-95 ℃) and pH value (pH 5.0, 6.0, 7.0, 8.0 and 9.0).
In contrast to other debilitating p53 mutations, the magni- tude of the molecular defect caused by the R337H mutation depends on pH. This observation can be put into perspective by studies of another tetramerization domain mutant, R337C. p53tet-R337H, with Tm values of 52 ℃ at pH 6 and 43 ℃ at pH 7 (Fig. 3a), is considerably more stable than p53tet-R337C, with a Tm value of 33 ℃ at pH 7 (ref. 8). However, these two mutants are similarly unstable at pH 8 (Fig. 3a). Therefore, intracellular pH is likely to modulate the function of p53-R337H, ranging from the observed wild type-like func- tion9,14 to dysfunction, as seen in adrenal tumor cells.
In p53tet-wt, the guanidinium moiety of Arg 337 participates in a salt bridge with the carboxylate group of Asp 352 (ref. 12) and contributes significantly to the stability of the tetrameriza- tion domain13. This salt bridge exists between pH 5 and 9 because the individual functional groups remain charged within this range (Arg has a pKa value of ~12; Asp, pKa value of 4) 17. The insensitivity of p53tet-wt denaturation to changes in pH (Fig. 3b) supports this analysis. Furthermore, the methylene groups of Arg 337 form a stable hydrophobic core with the aliphatic side chains of other residues, including Ile 332, Met 340, Phe 341 and Ile 348; these interactions contribute to p53tet- wt stability13. The R337H mutation alters both the charge and nonpolar characteristics at position 337. First, the His side chain is shorter and less hydrophobic than that of Arg. Second, the His side chain terminates with an imidazole ring that is positively charged only at pH values below its pKa value. The nominal pKa value of the His side chain is ~6.5; however, this can be influ- enced by chemical environment and could be shifted to a higher value by, for example, energetically favorable electrostatic inter- actions. Given these chemical differences between Arg and His, we hypothesize that the decreased stability of p53tet-R337H with increased pH is due to deprotonation of His 337 and the consequent loss of the stabilizing interhelix salt bridge.
His protonation and stability are correlated
The 1H NMR chemical shifts of nonexchangeable ring protons of His residues are sensitive to the protonation state of the imida- zole ring. For example, the chemical shift of the H2 proton in the protonated state is ~9 p.p.m., whereas this resonance shifts upfield to ~8 p.p.m. in the deprotonated state (Fig. 4a). 15N iso- tope-editing18 was used to selectively observe 1H resonances for 12C-bonded protons of uniformly 15N-labeled p53tet-R337H and p53tet-wt (Fig. 4b,c). One sharp resonance is observed between 8 and 9 p.p.m. for p53tet-wt (Fig. 4b), which corresponds to the single His residue in this construct. Two resonances are observed in this region for p53tet-R337H; one for that observed for p53tet-wt and a second for the H2 proton of His 337 (Fig. 4c). We measured the chemical shifts of imidazole protons in p53tet- R337H at different pH values at 37 °℃ (Fig. 4d) to determine the pKa of His 337 (Fig. 4e) and to relate the protonation state of this ring to the pH dependence of unfolding (at 37 ℃) for this can- cer-associated mutant. The pKa value for His 337 is 7.7 whereas that for the N-terminal His is 6.2. The latter pKa value is typical of solvent exposed His residues, whereas the value for His 337 is unusually high and reflects stabilization of the protonated state
a
1.0
37.º℃
Fraction denatured (fu)
pH 5
pH 6
PH 7
pH 8
Tm = 52 ℃
pH 9
0.5
1.0
f. at 37 ℃
T Tm = 32 ℃
0.5
0.0
0.0
5
6
7
8
9
pH
0
10
20
30
40
50
60
70
80
90
100
b
Temperature (ºC)
Fraction denatured (fu)
1.0
pH5
pH6
PH7
pH8
Tm =71 ℃
pH9
0.5
37 ℃
Tm = 76 ℃
0.0
0
10
20
30
40
50
60
70
80
90
100
Temperature (ºC)
of the imidazole ring by favorable electrostatic interaction with Asp 352.
The pH dependence of protonation of the His 337 side chain of p53tet-R337H coincides exactly with the pH dependence of domain unfolding. This is revealed by comparing the ring proto- nation versus pH at 37 ℃ (Fig. 4e) with the fraction of unfolded molecules versus pH at 37 ℃ (inset, Fig. 3a). The midpoints of these two curves coincide at ~pH 7.7, indicating that p53tet-R337H unfolds when His 337 is deprotonated. p53tet-R337H is more stable than p53tet-R337C when His 337 is protonated. Biochemical evidence suggests that when His 337 is protonated, p53-R337H functions like wild type p53. However, when the His 337-Asp 352 salt bridge is neutralized, p53tet- R337H unfolds at physiological temperature (37 °℃). Because proper folding of the tetramerization domain is required for p53 tumor suppressor function5, we suggest that R337H is a pH sen- sitive dysfunctional p53 mutant.
Implications
What is the relevance of our biophysical observations to the occurrence of ACC in Brazilian children that harbor p53-R337H? The genetic profiles of 35 of 36 pediatric ACC patients studied identify the mutation of Arg 337 to His is linked to cancer formation9. The R337H mutation increases the odds of developing ACC by ~500,000-fold (data not shown). Further supporting the contribution of the R337 mutation to ACC, the tumors from p53-R337H individuals frequently exhibit loss of heterozygosity by retaining only the mutant (R337H) allele, and the missense p53 protein is expressed at elevated levels in the nucleus9. Paradoxically, Ribeiro et al.9 and Lomax et al.14
npg
a
d
+
H-C-NH
H-C-N
His 337
N-term
His H2
4
3
4
CH 2
3
2
CH
H2
pH
C-NH
5
1
5
1
NH
5.1
CH2
CH2
5.6
6.2
ppm
~8 ppm
6.5
6.8
7.1
b
N-term
7.8
His H2
8.3
8.9
9.4
tunel
9.9
9.2
9.0
8.8
8.6
8.4
8.2
8.0
7.8
1H (ppm)
9.6
9.2
8.8
8.4
8.0
7.6
7.2
6.8
6.4
1H (ppm)
e
His 337
9.0
His 337
c
H2
W
pKa = 7.74
1H (ppm)
8.5
9.6
9.2
8.8
8.4
8.0
7.6
7.2
6.8
6.4
8.0
N-term His
1H (ppm)
pKa = 6.24
7.5
5
6
7
8
9
10
pH
explored the possibility that p53-R337H is functionally deficient with respect to wild type p53 but were unable to identify differ- ences between the two proteins by a variety of transfection pro- moter-reporter and growth suppression assays. These assays may not be sufficiently sensitive to detect subtle but biologically sig- nificant functional alterations19. Furthermore, the dependence of p53-R337H function on pH was not considered in these pre- vious studies. To properly address the consequences of the R337H mutation on tumor suppressor activity, more physiolog- ically relevant assays will need to be employed, such as the gener- ation of a ‘knock-in mutation’ mouse model and the establishment of adrenal tumor cell lines; these studies are cur- rently underway.
p53 is a modular protein whose function requires the tertiary and quaternary structural integrity of several domains, includ- ing the DNA binding and tetramerization domains4. The link between mutation-induced p53 tetramer instability and Li- Fraumeni and Li-Fraumeni-like syndromes has been established for p53-L344P and p53-R337C, respectively8. A feature that sets the ACC cases analyzed here apart from those associated with LFS and LFLS is the absence of other accompanying tumor types. In contrast to the broader range of tumors that develop in patients with LFS and LFLS germline p53 mutations, only ACC tumors appear in children that inherit the R337H mutation.
Because the R337H mutation introduces extraordinary pH sen- sitivity into p53, the origins of the tissue specificity observed in the ACC cases may be associated with an elevated pH within adrenal cells. The adrenal gland is known to undergo extensive cellular remodeling during pre- and postnatal development20-22, and this process requires the functional integrity of an apoptotic response. Under certain circumstances, apoptotic cells are known to have increased intracellular pH (~7.9)23. In the devel- oping adrenal gland, these factors may collaborate to create a cellular environment - characterized by an elevated pH - that destabilizes the tetramerization domain of p53-R337H, leading to loss of tumor suppressor function. For example, in cells with an intracellular pH of 7.9, the tetramerization domain of p53-R337H would be destabilized to the same extent as p53-R337C. In contrast, in cells with an intracellular pH closer to 7.0, p53-R337H may be sufficiently stable to function nor- mally. Localized p53 dysfunction in adrenal cells may allow the outgrowth of abnormal cells, which are normally controlled or eliminated through p53 tumor suppressor activity, to develop into tumors. Although speculative, this model is consistent with the exclusive association of the p53-R337H genotype with an extraordinarily high predisposition to ACC and with the R337H mutation introducing a pH-sensitive molecular switch that affects tetramerization domain stability. Our findings support a
npg
letters
novel mechanism of tumorigenesis, a pH-dependent p53 tumor suppressor dysfunction, that leads to tissue-specific tumor development.
Methods
Expression and purification of p53tet. cDNA segments for p53tet-wt (310-360) and p53tet-R337H (311-360) were amplified by PCR and subcloned into pET28a (Novagen). The proteins were expressed in Escherichia coli BL21(DE3) cells using standard proce- dures. The proteins were purified using Ni2+-affinity and cation- exchange chromatography (Amersham-Pharmacia resins). Between the two columns, the His-tag was removed by thrombin cleavage (leaving residues GSHM at the N-terminus). The identity of both proteins was confirmed using MALDI-TOF mass spectrometry.
Gel filtration chromatography. p53tet-wt and p53tet-R337H were resolved using a Superose-12 HR gel filtration column (Amersham-Pharmacia) in 20 mM NaH2PO4, pH 8, and 500 mM NaCI at 4 ℃. The retention times of mutant and wild type p53tet were compared to those of protein standards (BioRad).
CD studies. CD spectra were recorded using an AVIV model 62A DS CD spectropolarimeter. Samples of either p53tet-wt or p53tet- R337H at 10 uM were prepared in 20 mM NaH2PO4 and 250 mM NaCI at various pH values in 1 cm x 1 cm quartz cells. Thermal denat- uration curves were prepared by monitoring ellipticity at 222 nm as a function of temperature and converting these values into mole fraction of denatured protein (fu) assuming a two-state unfolding model.
NMR spectroscopy. We prepared 15N-labeled mutant and wild type p53tet for NMR studies as given above, except that E. coli BL21(DE3) cells were grown in defined media24 containing 15N-ammonium chloride. The p53tet samples for 2D-NMR experi- ments were at 1-2 mM in 20 mM NaH2PO4, 50 mM NaCI, 10% (v/v) D20, 0.2 mM sodium 3-(trimethylsilyl) propionate-2,2,3,3-d4 (TSP) (Cambridge Isotopes) and 0.02% (w/v) sodium azide, pH 5.0. The samples for 1D-NMR experiments were at 1-2 mM in 20 mM NaH2PO4, 250 mM NaCI, 10% (v/v) D20, 0.2 mM TSP and 0.02% (w/v) sodium azide at various pH values. All NMR spectra were acquired using a Varian Inova 600 MHz spectrometer (Varian NMR Systems) with a 5 mm triple-resonance probe equipped with x, y, z axis pulsed magnetic field gradients (PFGs). The 2D 1H-15N sensitivity- enhanced TROSY pulse sequence of Weigelt25, as implemented in Varian’s ProteinPack library, was used. Spectra were processed using the program Felix 98 (Molecular Simulations, Inc.). TSP was used as an internal standard for 1H chemical shift referencing. 15N chemical shifts were indirectly referenced using the ratio of 1H and 15N gyro- magnetic ratios18. A 15N filter sequence was used to selectively
observe 12C-bonded His H2 protons during pH titrations19. This con- sisted of a 90° 1H pulse followed by a 1/4JAN delay, 180° pulses to both 1H and 15N, and an additional 1/4JAN delay. Immediately before detection, a 90° 15N pulse was applied to convert antiphase 1H-15N magnetization into unobservable zero- and two-quantum magneti- zation. A selective 3-9-19 180° 1H pulse centered on water was used to refocus His H2 protons and not water. In addition, pulsed field gradients were applied before and after the 180° pulses to further suppress the water resonance.
Acknowledgments
The authors would like to thank C. Arrowsmith for providing a plasmid containing the wild type p53tet DNA sequence, and members of the Molecular Oncogenesis Program at St. Jude and the Kriwacki and Zambetti laboratories for stimulating discussion. This work was supported by the American Lebanese Syrian Associated Charities, the American Cancer Society, the NCI and a Cancer Center (CORE) Support Grant.
Correspondence should be addressed to R.W.K. email: richard.kriwacki@stjude.org
Received 20 August, 2001; accepted 24 November, 2001.
1. Prives, C. & Hall, P.A. J. Pathol. 187, 112-126 (1999).
2. Levine, A.J. Cell 88, 323-331 (1997).
3. Hollstein, M., Sidransky, D., Vogelstein, B. & Harris, C.C. Science 253, 49-53 (1991).
4. Pavletich, N.P., Chambers, K.A. & Pabo, C.O. Genes Dev. 7, 2556-2564 (1993).
5. Stürzbecher, H.W. et al. Oncogene 7, 1513-1523 (1992).
6. Varley, J.M. et al. Oncogene 12, 2437-2442 (1996).
7. Lomax, M.E. et al. Oncogene 14, 1869-1874 (1997).
8. Davison, T.S., Yin, P., Nie, E., Kay, C. & Arrowsmith, C.H. Oncogene 17, 651-656 (1998).
9. Ribeiro, R.C. et al. Proc. Natl. Acad. Sci. USA 98, 9330-9335 (2001).
10. Lee, W. et al. Nature Struct. Biol. 1, 877-890 (1994).
11. Clore, G.M. et al. Nature Struct. Biol. 2, 321-333 (1995).
12. Jeffrey, P.D., Gorina, S. & Pavletich, N.P. Science 267, 1498-1502 (1995).
13. Mateu, M.G. & Fersht, A.R. EMBO J. 17, 2748-2758 (1998).
14. Lomax, M.E., Barnes, D.M., Hupp, T.R., Picksley, S.M. & Camplejohn, R.S. Oncogene 17, 643-649 (1998).
15. Pervushin, K., Riek, R., Wider, G. & Wuthrich, K. Proc. Natl. Acad. Sci. USA 94, 12366-12371 (1997).
16. Mathews, C.K. & van Holde, K.E. Biochemistry (Benjanim Cummings Publishing Co., New York; 1996).
17. Creighton, T.E. Proteins: structures and molecular properties (Freeman & Co., New York; 1993).
18. Cavanagh, J., Fairbrother, W.J., Palmer, A.G. III & Skelton, N.J. Protein NMR spectroscopy (Academic Press, New York; 1996).
19. Chao C., Saito S., Anderson C.W., Appella, E. & Xu, Y. Proc. Natl. Acad. Sci. USA 97, 11936-11941 (2000).
20. Mesiano, S. & Jaffe, R.B. Endocr. Rev. 18, 378-403 (1997).
21. Jaffe, R.B. et al. Endocr. Res. 24, 919-926 (1998).
22. Spencer, S.J., Mesiano, S., Lee, J.Y. & Jaffe, R.B. J. Clin. Endocrinol. Metab. 84, 1110-1115 (1999).
23. Dai, H.Y., Tsao, N., Leung, W.C. & Lei, H.Y. Radiat. Res. 150, 183-189 (1998).
24. Neidhardt, F.C., Bloch, P.L. & Smith, D.F. J. Bacteriol. 119, 736-747 (1974).
25. Weigelt, J. J. Am. Chem. Soc. 120, 10778-10779 (1998).